RESUMEN
OBJECTIVE: To describe the diagnostic yields of test strategies with and without fiberoptic bronchoscopy and bronchoalveolar lavage (FO-BAL), as well as outcomes, in cancer patients with acute respiratory failure (ARF). DESIGN: Prospective observational study. SETTING: Fifteen intensive care units in France. PATIENTS: In all, 148 cancer patients, including 45 bone marrow transplant recipients (27 allogeneic, 18 autologous) with hypoxemic ARF. INTERVENTION: None. RESULTS: Overall, 146 causes of ARF were identified in 128 patients (97 [66.4%] pulmonary infections). The cause of ARF was identified in 50.5% of the 101 patients who underwent FO-BAL and in 66.7% of the other patients. FO-BAL was the only conclusive test in 34 (33.7%) of the 101 investigated patients. Respiratory status deterioration after FO-BAL occurred in 22 of 45 (48.9%) nonintubated patients, including 16 (35.5%) patients who required ventilatory support. Hospital mortality was 55.4% (82 deaths) overall and was not significantly different in the groups with and without FO-BAL. By multivariate analysis, mortality was affected by characteristics of the malignancy (remission, allogeneic bone marrow transplantation), cause of ARF (ARF during neutropenia recovery, cause not identified), and need for life-sustaining treatments (mechanical ventilation and vasopressors). CONCLUSION: In critically ill cancer patients with ARF, a diagnostic strategy that does not include FO-BAL may be as effective as FO-BAL without exposing the patients to respiratory status deterioration.